Clinical and dermoscopic characteristics of melanomas on nonfacial chronically sun-damaged skin

ORIGINAL

ARTICLE

Clinical and dermoscopic characteristics of melanomas on nonfacial chronically sun-damaged skin Natalia Jaimes, MD,a Ashfaq A. Marghoob, MD,b Harold Rabinovitz, MD,c Ralph P. Braun, MD,d Alan Cameron, MBBS,e Cliff Rosendahl, MBBS, PhD,e Greg Canning, MBBS,f and Jeffrey Keir, MBBS, MFamMedg Medell
ın, Colombia; New York, New York; Plantation, Florida; Z€ urich, Switzerland; and Brisbane, Townsville, and Ballina, Australia Background: Melanomas on chronically sun-damaged skin (CSDS) can be difficult to identify and often manifest morphologic features that overlap with benign lesions. Objective: We describe and analyze the clinical and dermoscopic characteristics of melanomas on nonfacial CSDS. Methods: Melanoma cases on nonfacial CSDS were retrospectively identified from the biopsy specimen logs of 6 melanoma clinics. Clinical and dermoscopic images were combined into 1 database. Demographics, clinical, dermoscopic, and histopathologic information were analyzed. Descriptive frequencies were calculated. Results: One hundred eighty-six cases met the inclusion criteria: 142 melanomas in situ (76%) and 39 invasive (21%; mean thickness, 0.49 mm). Lentigo maligna was the most common histopathologic subtype (n = 76; 40.9%). The most frequent dermoscopic structures were granularity (n = 126; 67.7%) and angulated lines (n = 82; 44%). Vascular structures were more frequent in invasive melanomas (56% vs 12% of in situ melanomas). Most manifested 1 of 3 dermoscopic patterns: patchy peripheral pigmented islands, angulated lines, and tan structureless with granularity pattern. Limitations: This was a retrospective study, and evaluators were not blinded to the diagnosis. In addition, interobserver concordance and sensitivity and specificity for dermoscopic structures were not evaluated. Conclusion: Outlier lesions manifesting dermoscopic structures, such as granularity, angulated lines, or vessels and any of the 3 described dermoscopic patterns should raise suspicion for melanoma. ( J Am Acad Dermatol http://dx.doi.org/10.1016/j.jaad.2015.02.1117.) Key words: actinic damage; dermoscopy; lentigo maligna; melanoma; sun-damaged skin; ultraviolet radiation.

I

dentifying melanomas located on chronically sun-damaged skin (CSDS) is challenging because their characteristics often clinically overlap with benign lesions (Fig 1). Dermoscopy has proven useful in detecting and differentiating melanoma from many benign lesions, including

From the Dermatology Service,a Aurora Skin Cancer Center and Universidad Pontificia Bolivariana, Medell
ın; Dermatology Service,b Memorial Sloan-Kettering Cancer Center, New York; Skin and Cancer Associates,c Plantation; Department of Dermaurich; School of Medicine,e The tology,d University Hospital Z€ University of Queensland, Brisbane; Hermit Park Clinic and Skin Cancer Care,f Townsville; and the Northern Rivers Skin Cancer Clinic,g Ballina. Funding sources: None.

Abbreviations used: CSDS: SSM: LM: PPPI:

chronically sun-damaged skin superficial spreading melanoma lentigo maligna patchy peripheral pigmented islands

Conflicts of interest: None declared. Accepted for publication February 16, 2015. Correspondence to: Jeffrey Keir, MBBS, MFamMed, PO Box 892, Ballina, NSW 2478, Australia. E-mail: [email protected]. Published online March 27, 2015. 0190-9622/$36.00 Ó 2015 by the American Academy of Dermatology, Inc. http://dx.doi.org/10.1016/j.jaad.2015.02.1117

1

J AM ACAD DERMATOL

2 Jaimes et al

n 2015

by pathology, melanomas from which previous solar lentigo and seborrheic keratosis.1-3 The dermobiopsy specimens had been obtained, and recurrent scopic attributes associated with melanomas on melanomas were excluded. facial CSDS, primarily lentigo maligna (LM), Two authors evaluated the clinical and dermoare well described. These descriptions may be scopic characteristics of all lesions (N.J. and A.A.M.). dependent on the specific histologic structure Information collected for each case included age, of the facial skin (ie, closely packed pilosebaceous sex, and histopathologic (ie, Breslow thickness, units, absence of rete ridges). The differing subtype, and solar elastosis), structure of nonfacial CSDS clinical (ie, location, size, co(ie, epidermal atrophy and CAPSULE SUMMARY lor, borders, outlier lesion, variable flattening of the derand background skin), and moepidermal junction) may Dermoscopic attributes associated with dermoscopic characteristics. affect the appearance of LM amelanotic melanoma, superficial Clinical colors included light and other melanoma subspreading melanoma, lentigo maligna, brown, dark brown, bluetypes on patients with acral lentiginous melanoma, and nodular gray, white, pink, and black. nonfacial CSDS, and their melanoma are well described. Background skin was evaludermoscopic features remain We report the clinical and dermoscopic ated for signs suggestive of to be clearly defined. In this characteristics of melanomas on CSDS, such as excessive and/ case series, we describe and nonfacial chronically sun-damaged skin. or large lentigines, and for analyze the clinical and derthe presence of other benign moscopic characteristics of These findings will improve the early lesions. Solar lentigines sur186 melanomas located on detection of melanomas on nonfacial rounding the melanoma nonfacial CSDS with the aim chronically sun-damaged skin. (evaluated from the overof improving knowledge of view images) were quantithe clinical and dermoscopic fied and graded into few (\20), moderate (20-100), characteristics of these melanomas. or severe ([100) in number. Dermoscopic images were captured with polarized METHODS and/or nonpolarized dermoscopy and were assessed Cases of melanomas on nonfacial CSDS skin were for the presence or absence of previously described retrospectively identified from the biopsy specimen melanoma-specific structures4 (Supplemental log and image database of 6 dermatology and Table I). The presence or absence of angulated lines primary care skin cancer clinics (Memorial Sloanand a blue-white veil required uniform consensus of 3 Kettering Cancer Center, New York, NY; Skin and authors (N.J., A.A.M., and H.R.). Descriptive freCancer Associates, Plantation, FL; Melanoma quencies were reported. Signature Skin Cancer Centre, Brisbane, Australia; Northern Rivers Skin Cancer Clinic, Ballina, Australia; Beenleigh Family Practice, Brisbane, Australia; and Hermit Park Clinic and Skin Cancer RESULTS Care, Townsville, Australia). This study was conductOf 186 cases that met the inclusion criteria, 112 ed in accordance with the institutional review board (60%) patients were male and 62 (33%) were female. at Memorial Sloan-Kettering Cancer Center. The average patient age was 68.5 years (range, 37-93 Nonfacial CSDS was defined as any anatomic site years). There were 142 (76.3%) in situ and 39 (21%) that fulfilled 1 of the following characteristics: invasive melanomas (average thickness, 0.49 mm location on the upper aspect of the back, chest, or [range, 0.12-1.6 mm]). LM was the most common upper extremities; location on other sites, such as the histopathologic subtype (n = 76; 40.9%), followed by lower aspect of the back, abdomen, legs and thighs superficial spreading melanoma (SSM; n = 42; with signs of sun damage, such as moderate to severe 22.6%). Twenty-one cases (11.3%) revealed a comnumbers of lentigines or at least a moderate degree bination of 2 subtypes of melanoma (Table I). of solar elastosis present in the skin biopsy Lesions were located on the back (n = 89; 47.8%), specimen. upper extremities (n = 61; 32.8%), chest and lower Inclusion criteria consisted of biopsy-proven extremities (n = 16; 8.6% each), and abdomen (n = 4; melanomas located on nonfacial CSDS with high 2.2%). The majority were outlier lesions (n = 142; quality clinical and dermoscopic images. Melanomas 76.3%), with poorly defined borders (n = 126; 67.7%) on the head, neck, volar surface, breasts, genitalia, and an average largest diameter of 9.44 mm (range, and buttocks were excluded. In addition, mela3-40 mm). The background skin surrounding each nomas arising in association with a nevus proven melanoma revealed signs of CSDS, including solar d

d

d

J AM ACAD DERMATOL

Jaimes et al 3

VOLUME jj, NUMBER j

Table I. Histopathologic characteristics of melanomas on nonfacial chronically sun-damaged skin (n = 186)* Histopathologic characteristics

Fig 1. Severe chronically sun-damaged skin on the back of a 62-year-old man.

lentigines (n = 184; 98.9%), which were moderate to severe in quantity in 146 (78%) cases (Table II). Clinically, lesions revealed multiple colors, most frequently light- and dark-brown (n = 161 [86.6%] and n = 158 [84.9%], respectively), followed by pink (n = 56; 30%), blue-gray (n = 43; 23%), and white (n = 10; 5.4%). Black color was only appreciated in 2 cases (1%), both invasive. Two colors were seen in 61% (n = 113), 3 colors in 19.9% (n = 37), and $4 colors in 9.7% (n = 18). Invasive melanomas displayed [4 colors compared with in situ melanomas (10/39 [25.7%] vs 5/142 [3.5%]), in particular pink (21/39 [53.8%] vs 31/142 [21.8%]), blue-gray (15/39 [38.5%] vs 25/142 [17.6%]), white (7/39 [17.9%] vs 3/142 [2.1%]), and black (2/39 [5.1%] vs 0%; Table II). Colors were better appreciated under dermoscopy. Blue-gray and white were more conspicuous under nonpolarized dermoscopy (Tables II and III). The most frequent structures seen were granularity (n = 126; 67.7%), angulated lines (n = 82; 44.1%), and atypical dots (rarely globules; n = 68; 36.6%). Granularity was seen in association with pigmented structures (ie, pigment network or structureless brown areas) in 124 (67%) cases. Other structures observed were atypical pigment network (n = 55; 29.6%), asymmetrically pigmented perifollicular/adnexal openings (n = 51; 27.4%), peripheral structureless tan areas (n = 46; 24.7%), and scar-like areas (n = 37; 19.9%). Vascular structures were observed in\33% of cases, particularly milky red areas (n = 46; 24.7%) and dotted vessels (n = 43; 23%; Table III). Some dermoscopic structures were more frequently seen in invasive melanomas than in situ melanomas. These included negative network (6/39 [15.4%] vs 2/142

Melanoma in situ Invasive melanoma Missing data Breslow thickness, mm, mean (range) Mitosis, mean (range) Melanoma subtype LM SSM Lentiginous Mixed desmoplastic MM Nevoid melanoma Missing data/not specified Two melanoma subtypes LM 1 nevoid MM LM 1 SSM Lentiginous 1 nevoid MM Lentiginous 1 SSM Solar elastosis Present Absent Not reported/not available

Total (n) Percentage

142 76.3% 39 21.0% 5 2.7% 0.49 (0.12-1.6) 0.30 (0-1) 76 43 19 1 3 24 21 8 5 5 3

40.9% 22.6% 10.2% 0.5% 1.6% 12.9% 11.3% 4.3% 2.7% 2.7% 1.6%

42 2 142

22.6% 1.1% 76.3%

LM, Lentigo maligna; SSM, superficial spreading melanoma. *Pathology reports incomplete for Breslow thickness and/or subtype in 5 cases.

[1.4%]), scar-like areas (14/39 [35.9%] vs 21/142 [14.8%]), crystalline structures (12/39 [30.8%] vs 11/ 142 [7.7%]), blue-white veil (3/39 [7.7%] vs 0), offcentered blotch (3/39 [7.7%] vs 4/142 [2.8%]), and vascular structures, such as dotted vessels (22/39 [56.4%] vs 17/142 [12%]), serpentine vessels (17/39 [43.6%] vs 9/142 [6.3%]), polymorphous vessels (ie, dotted and serpentine; 13/39 [33.3%] vs 8/142 [5.6%]), milky red area/vascular blush (21/39 [53.8%] vs 21/142 [14.8%]), and red globules (2/39 [5%] vs 0%; Table III). One of 3 specific dermoscopic patterns was seen in 78% (n = 146) of cases (Table IV). Patterns were defined as specific combinations of dermoscopic structures covering [50% of a lesion that were consistently observed in a substantial number of lesions. LM was the most common histopathologic subtype, comprising 40% to 45% of each group. The most frequent pattern (n = 70; 37.6%) was focal islands of pigmentation (network or structureless areas) located towards the periphery (Fig 2). Toward the center of these melanomas were focal featureless areas, most often scar-like or hypopigmented. Of these, 13 (7%) had crystalline structures, and 21 (11.3%) revealed vascular blush, most commonly within the focal hypopigmented areas. This pattern, named ‘‘patchy peripheral pigmented

J AM ACAD DERMATOL

4 Jaimes et al

n 2015

Table II. Clinical characteristics of melanomas on nonfacial chronically sun-damaged skin (n = 186) Overall (N = 186) Clinical characteristics

Location Back Chest Abdomen Upper extremities Lower extremities Size (mm), mean (range) Colors 1 color 2 colors 3 colors 4 colors 5 colors Light brown Dark brown Pink Blue-gray White Black Borders Poorly defined Well-defined, irregular Well-defined, regular Background skin Lentigines Few Moderate Severe Guttate hypomelanosis Seborrheic keratosis Cherry hemangiomas Outlier lesion Yes No

Total (n)

%

89 47.8 16 8.6 4 2.2 61 32.8 16 8.6 9.44 (3-40)

MM in situ (N = 142) Total (n)

%

69 48.6 15 10.6 2 1.4 47 33.1 9 6.3 9.02 (3-20)

Invasive melanoma (N = 39) Total (n)

%

20 1 1 13 4

51.3 2.6 2.6 33.3 10.2 9.33 (4-22)

18 113 37 15 3 161 158 56 43 10 2

9.7 60.8 19.9 8.1 1.6 86.6 84.9 30.1 23.1 5.4 1.1

14 97 26 3 2 126 124 31 25 3 0

9.9 68.3 18.3 2.1 1.4 88.7 87.3 21.8 17.6 2.1 0

4 16 9 9 1 30 29 21 15 7 2

10.2 41.0 23.1 23.1 2.6 76.9 74.4 53.8 38.5 17.9 5.1

126 44 16

67.7 23.7 8.6

94 35 13

66.2 24.6 9.2

29 7 3

74.4 17.9 7.7

184 38 79 67 173 76 30

98.9 20.4 42.5 36.0 93.0 40.9 16.1

141 30 61 50 132 51 17

99.3 21.1 43.0 35.2 93.0 35.9 12.0

38 7 15 16 37 24 12

97.4 17.9 38.5 41.0 94.9 61.5 30.8

142 44

76.3 23.7

103 39

72.5 27.5

34 5

87.2 12.8

MM, Malignant melanoma.

islands’’ (PPPI), was most frequently seen on the back (n = 43; 61.4%). These lesions had the greatest average size (10.72 mm; Table IV). The second most common dermoscopic pattern (n = 56; 30.1%) was termed ‘‘angulated line pattern’’ (Fig 3); this was comprised of angulated lines with or without the presence of granularity or circle within a circle and was most commonly seen on the upper extremities (n = 28; 50%). The third pattern, found in 20 (10.8%) cases, consisted of lesions with tan structureless areas and granularity (Fig 4). These lesions had the smallest average diameter (7.39 mm) and occurred most frequently on the back (n = 10; 50%; Table IV). The remaining 40 (21.5%) lesions did not adhere to any of the aforementioned patterns but had the following structures: granularity (14/40; 35%),

atypical network (11/40; 27.5%), peripheral tan structureless areas (12/40, 30%), atypical aggregated dots/globules (10/40, 25%), dotted vessels (10/40, 25%), milky red areas/vascular blush (5/40; 12.5%), serpentine vessels (4/40, 10%), asymmetric perifollicular hyperpigmentation (7/40; 17.5%), angulated lines (6/40, 15%), crystalline structures (5/40; 12.5%), and scar-like areas, negative network, off-centered blotch, and blue-white veil (1/40; 2.5% each). Melanomas with a nonspecific pattern were more frequently seen on the back (n = 20; 50%) and had an average diameter of 9.40 mm (Table IV).

DISCUSSION Melanomas occurring on nonfacial CSDS can be challenging to detect because they can mimic lentigines or nevi and be camouflaged amongst

J AM ACAD DERMATOL

Jaimes et al 5

VOLUME jj, NUMBER j

Table III. Dermoscopic characteristics of melanomas on nonfacial chronically sun-damaged skin (n = 186)* Overall (N = 186) Dermoscopic characteristics

Regression structures Granularity or peppering Scar-like areas Angulated lines Aggregated dots (rarely globules) Atypical pigment network Asymmetric perifollicular hyperpigmentation Peripheral tan structureless areas Vascular structures Dotted vessels Serpentine vessels Polymorphous (dotted 1 serpentine) Milky red areas/vascular blush Red globules White shiny structures Crystalline structures White shiny areas Rosettes Negative network Circle within circle Off-center blotch Blue-white veil Streaks Pseudopods Radial projections Colors seen under dermoscopy Light brown Dark brown Blue-gray Blue-white Pink White Overall dermoscopic appearance/pattern Patchy peripheral pigmented islandsy Angulated line pattern Tan structureless and granularity pattern None of the above

Total (n)

%

126 37 82 68

67.7 19.9 44.1 36.6

55 51

MM in situ (N = 142) Total (n)

Invasive melanoma (N = 39)

%

Total (n)

%

98 21 64 48

69.0 14.8 45.1 33.8

25 14 15 17

64.1 35.9 38.5 43.6

29.6 27.4

41 42

28.9 29.6

9 8

23.1 20.5

46

24.7

34

23.9

10

25.6

43 26 21

23.1 14.0 11.3

17 9 8

12.0 6.3 5.6

22 17 13

56.4 43.6 33.3

46

24.7

21

14.8

21

53.8

2

1.1

0

0

2

5.1

26 4 3 10 12 7 3

14.0 2.2 1.6 5.4 6.5 3.8 1.6

11 1 2 2 9 4 0

7.7 0.7 1.4 1.4 6.3 2.8 0

12 2 1 6 3 3 3

30.8 5.1 2.6 15.4 7.7 7.7 7.7

1 0

0.5 0

1 0

0.7 0

0 0

0 0

179 170 129 65 55 52

96.2 91.4 69.4 34.9 29.6 28.0

137 130 94 42 26 27

96.5 91.5 66.2 29.6 18.3 19.0

38 36 31 21 26 22

97.4 92.3 79.5 53.8 66.7 56.4

70

37.6

52

36.6

15

3

56 20

30.1 10.8

44 15

31.0 10.6

11 5

28.2 12.8

40

21.5

31

21.8

8

20.5

MM, Malignant melanoma. *Pathology reports incomplete for Breslow thickness in 5 cases. y Crystalline structures were centrally located within the focal hypopigmented areas in 13 cases (7%): 6 melanomas in situ (4.2%) and 6 invasive melanomas (15.4%). Vascular blush was centrally located within the focal hypopigmented areas in 21 cases (11.3%): 13 melanomas in situ (9.2%) and 8 invasive melanomas (20.5%).

many surrounding benign lesions. Careful clinical examination together with dermoscopy can assist in the evaluation of these patients. There have been few studies examining these lesions.5,6

Traditional melanoma classification is based on clinical and pathologic characteristics and includes 4 melanoma subtypes: SSM, LM, nodular melanoma, and acral lentiginous melanoma.7,8 Recently, the

J AM ACAD DERMATOL

6 Jaimes et al

n 2015

Table IV. Dermoscopic patterns seen in melanomas on nonfacial chronically sun-damaged skin (n = 186) Patchy peripheral pigmented islands, n = 70 (37.6%)

Angulated line pattern, n = 56 (30.1%)

Tan structureless and granularity pattern, n = 20 (10.8%)

No pattern, n = 40 (21.5%)

Schematic

Anatomic site Back Chest Upper extremities Lower legs Abdomen MM subtype LM SSM Lentiginous Mixed desmoplastic Nevoid melanoma Not specified Two MM subtypes LM 1 nevoid LM 1 SSM Lentigines 1 nevoid Lentigines 1 SSM Average size (range), mm

43 5 16 4 2

(61.4) (7.1) (22.9) (5.7) (2.9)

16 5 28 5 2

(28.6) (8.9) (50.0) (8.9) (3.6)

10 2 5 3 0

(50.0) (10.0) (25.0) (15.0)

20 4 12 4 0

(50.0) (10.0) (30.0) (10.0)

28 18 8 0 1 8 7 1 3 1 2 10.72 mm

(40.0) (25.7) (11.4)

23 11 4 0 0 10 8 3 2 2 1 8.55 mm

(41.1) (19.6) (7.1)

9 4 1 0 0 4 2 2 0 0 0 7.39 mm

(45.0) (20.0) (5.0)

17 8 6 1 2 2 4 2 0 2 0 9.40 mm

(42.5) (20.0) (15.0) (2.5) (5.0) (5.0) (10.0) (5.0)

(1.4) (11.4) (10.0) (1.4) (4.3) (1.4) (2.9) (4-40)

(17.9) (14.3) (5.4) (3.6) (3.6) (1.8) (5-20)

(10.0) (10.0) (10.0)

(3-20)

(5.0) (3-22)

LM, Lentigo maligna; MM, malignant melanoma; SSM, superficial spreading melanoma.

Fig 2. Peripheral pigmented islands pattern. Melanoma (Breslow thickness 0.37 mm) on the abdomen of a 79year-old man. Dermoscopic evaluation revealed atypical network islands located at the periphery of the lesion in a disorganized fashion, intercalated with hypopigmented areas and a central scar-like area.

recognition of distinct patterns of genetic alterations led to a proposed classification based on anatomic location and degree of sun exposure.9 These groups included melanomas on CSDS, intermittently sunexposed skin, and minimally or nonesun-exposed skin (ie, acral and mucosal melanomas). Melanomas on intermittently sun-exposed skin have been

Fig 3. Angulated line pattern. Melanoma in situ on the upper extremity of a 67-year-old woman. Dermoscopy revealed angulated lines consisting of multiple instances of grey or brown straight lines or pigment borders meeting at angles over a light-brown background.

associated with SSM, a large number of nevi, and BRAF mutations in #60% of cases. In contrast, melanomas on CSDS have been associated with continuous sun-exposure, LM and, in 30% of cases, c-KIT mutations.9-12 CSDS has been correlated not only with LM, but also with elastosis, actinic keratoses, and number of solar lentigines, suggested to be one of the strongest predictors for melanoma,

J AM ACAD DERMATOL

VOLUME jj, NUMBER j

Fig 4. Tan structureless and granularity pattern. Melanoma in situ on the back of a 78-year-old man. Dermoscopy revealed granularity over a tan structureless background.

particularly LM (odds ratio, 15.93; P \ .001).10-13 Similarly, our study showed that the largest group of melanomas located on nonfacial CSDS were of the LM type (40.9%) and were associated with a background of lentigines (98.9%). The second most frequent type of melanoma on nonfacial CSDS was SSM (22.6%), followed by lentiginous melanoma (10.2%; Table II). The lentiginous type of melanoma is considered by some to represent a distinct subset of melanomas.14,15 Although most melanomas were relatively small (average, 9.44 mm), 76.3% of the cases were outlier lesions, which facilitated their selection for further evaluation with dermoscopy and biopsy. In contrast to descriptions of benign nevi, melanomas on CSDS do not adhere to any of the benign dermoscopic patterns. The most frequent dermoscopic structure observed in these melanomas was granularity (67.7%), also known as peppering or multiple blue-gray dots, which was shown to correlate with regression and with melanin structures (free or within melanophages) in the papillary dermis.16 In melanocytic lesions, granularity is associated with a high sensitivity for melanoma (85%), a specificity of 93% to 99%, and an odds ratio of 3.5.16,17 Whether it correlates with thinner or thicker melanomas is still controversial. Seidenari et al18 reported that granularity was observed in both in situ and invasive melanomas in similar percentages, but Braun et al16 reported that this structure was primarily seen in high-grade dysplastic nevi, in situ, early invasive melanomas, and sun-damaged skin. Similarly, our study noted that granularity was found in both in situ (69%) and early invasive melanomas (64.1%). Follicular obliteration by pigment, a feature seen in invasive facial LM,5 was not noted in the series. Another frequent dermoscopic structure was angulated lines (44.1%), a new term proposed and

Jaimes et al 7

used in this study with the aim of unifying differing morphologies of previously described structures, including rhomboidal structures,19 zig zag lines,20 and polygonal structures.6 Rhomboidal structures were first described by Schiffner et al19 as part of a progression model for facial LM, consisting of small, straight-edged, dark-grey or dark-brown geometric shapes tightly centered on a follicle. Zig zag lines were first described by Slutsky and Marghoob20 as brown to bluish gray dots and lines arranged in an angulated linear pattern in facial LM and were later termed polygonal lines by Jaimes and Marghoob.4 Large polygons were first described by Keir6 as a frequent finding in lentiginous growth pattern melanomas on nonfacial nonacral skin, defined as complete or incomplete polygonal structures larger than facial rhomboidal structures, formed by straight gray or brown lines or straight edges of pigmented areas. It is not known whether all angulated line features represent the same biologic or pathological processes, or how distinct the different appearances are to different observers; therefore, we have used the term ‘‘angulated lines’’ to encompass all of the formerly described straight-sided geometric structures. Angulated lines are defined as multiple instances of grey or brown straight lines or straight pigment borders meeting at angles. Vascular structures were observed in \33% of cases, but were more frequently seen in invasive melanomas. Atypical vessels, such as dotted vessels, were present in 22 (56.4%) of the 39 invasive melanomas, compared to 17 (12%) of the in situ melanomas. Other vessel types were also more frequent in invasive melanomas, as were nonvascular features of negative network, crystalline structures, a blue-white veil, and scar-like areas (Table IV). Interestingly, 3 distinctive dermoscopic patterns not described in melanomas on facial CSDS were repeatedly seen in 146 (78%) of these melanomas on nonfacial CSDS (Table IV). The most frequent pattern (n = 70; 37.6%), was termed patchy peripheral pigmented islands and most commonly seen on the back. The second most frequent pattern (n = 50; 30.1%) was the angulated line pattern, more commonly observed on the upper extremities. The third pattern consisted of tan structureless areas and granularity, comparable to patterns seen in lichen planuselike keratosis and seen more frequently on the back. None of these patterns were manifest in 40 (22%) of lesions, which did, however, reveal dermoscopic structures not associated with benign nevi. Whether these patterns represent a continuum remains to be elucidated. It might be suggested that some of the lesions displaying a tan structureless and granularity

8 Jaimes et al

pattern or none of the 3 patterns may evolve into a PPPI or an angulated line pattern. Our study has limitations, including that evaluators were not blinded to the diagnosis, structures were identified by consensus and interobserver concordance, and that sensitivity and specificity were not evaluated. In addition, it was a retrospective study and diagnoses were based on original histopathology reports without pathology review. In conclusion, identifying clinical outlier lesions and routinely performing dermoscopy on these may aid in the recognition and differentiation of melanomas on nonfacial CSDS. Melanomas on CSDS may share some similar dermoscopic features with melanomas located on facial skin; structures most frequently observed in melanomas on CSDS include granularity, angulated lines, and atypical aggregated dots. Atypical pigment network, asymmetric perifollicular hyperpigmentation, peripheral structureless tan areas, and scar-like areas can also be seen. Vascular structures may be present, particularly in invasive melanomas. Structures such as negative network, aggregated dots/globules, crystalline structures, a blue-white veil, and scar-like areas are frequently observed in early invasive melanomas on nonfacial CSDS. In addition, [75% of melanomas located on nonfacial CSDS may feature 1 of 2 dermoscopic patterns: PPPI, tan structureless and granularity pattern, and angulated line pattern, the first 2 patterns are frequently found on the back and the last on upper extremities. The relationship of these patterns to the specific anatomy of nonfacial CSDS remains to be elucidated, but recognition of these patterns may raise suspicion for melanomas on CSDS and prompt the clinician to obtain a biopsy specimen. REFERENCES 1. Bafounta ML, Beauchet A, Aegerter P, Saiag P. Is dermoscopy (epiluminescence microscopy) useful for the diagnosis of melanoma? Results of a meta-analysis using techniques adapted to the evaluation of diagnostic tests. Arch Dermatol. 2001;137:1343-1350. 2. Kittler H, Pehamberger H, Wolff K, Binder M. Diagnostic accuracy of dermoscopy. Lancet Oncol. 2002;3:159-165.

J AM ACAD DERMATOL

n 2015

3. Benvenuto-Andrade C, Marghoob AA. Ten reasons why dermoscopy is beneficial for the evaluation of skin lesions. Exp Rev Dermatol. 2006;1:369-374. 4. Jaimes N, Marghoob AA. Overview of dermoscopy. In: Corona R, Tsao H, eds. UpToDate. Waltham (MA): Wolters Kluwer Health; 2014. 5. Lau YN, Affleck AG, Fleming CJ. Dermatoscopic features of extrafacial lentigo maligna. Clin Exp Dermatol. 2013;38:612-616. 6. Keir J. Dermatoscopic features of cutaneous non-facial non-acral lentiginous growth pattern melanomas. Dermatol Pract Concept. 2014;4:77-82. 7. Clark WH Jr, From L, Bernardino EA, Mihm MC. The histogenesis and biologic behavior of primary human malignant melanomas of the skin. Cancer Res. 1969;29:705-727. 8. McGovern VJ, Mihm MC Jr, Bailly C, et al. The classification of malignant melanoma and its histologic reporting. Cancer. 1973;32:1446-1457. 9. Curtin JA, Fridlyand J, Kageshita T, et al. Distinct sets of genetic alterations in melanoma. N Engl J Med. 2005;353:2135-2147. 10. Gandini S, Sera F, Cattaruzza MS, et al. Meta-analysis of risk factors for cutaneous melanoma: III. Family history, actinic damage and phenotypic factors. Eur J Cancer. 2005;41:2040-2059. 11. Kvaskoff M, Siskind V, Green AC. Risk factors for lentigo maligna melanoma compared with superficial spreading melanoma: a case-control study in Australia. Arch Dermatol. 2012;148:164-170. 12. Whiteman DC, Watt P, Purdie DM, et al. Melanocytic nevi, solar keratoses, and divergent pathways to cutaneous melanoma. J Natl Cancer Inst. 2003;95:806-812. 13. Thomas NE, Kricker A, From L, et al. Associations of cumulative sun exposure and phenotypic characteristics with histologic solar elastosis. Cancer Epidemiol Biomarkers Prev. 2010;19:2932-2941. 14. King R, Page RN, Googe PB, Mihm MC Jr. Lentiginous melanoma: a histologic pattern of melanoma to be distinguished from lentiginous nevus. Mod Pathol. 2005;18:1397-1401. 15. King R. Lentiginous melanoma. Arch Pathol Lab Med. 2011;135: 337-341. 16. Braun RP, Gaide O, Oliviero M, et al. The significance of multiple blue-grey dots (granularity) for the dermoscopic diagnosis of melanoma. Br J Dermatol. 2007;157:907-913. 17. Menzies SW, Kreusch J, Byth K, et al. Dermoscopic evaluation of amelanotic and hypomelanotic melanoma. Arch Dermatol. 2008;144:1120-1127. 18. Seidenari S, Ferrari C, Borsari S, et al. Reticular grey-blue areas of regression as a dermoscopic marker of melanoma in situ. Br J Dermatol. 2010;163:302-309. 19. Schiffner R, Schiffner-Rohe J, Vogt T, et al. Improvement of early recognition of lentigo maligna using dermatoscopy. J Am Acad Dermatol. 2000;42:25-32. 20. Slutsky JB, Marghoob AA. The zig-zag pattern of lentigo maligna. Arch Dermatol. 2010;146:1444. 21. Cognetta AB, Stolz W, Katz B, Tullos J, Gossain S. Dermatoscopy of lentigo maligna. Dermatol Clin. 2001;19:307-318.

J AM ACAD DERMATOL

Jaimes et al 9

VOLUME jj, NUMBER j

Supplemental Table I. Melanoma-specific structures4,6,19-21 Melanoma-specific structure

Definition

Atypical pigment network Negative network Atypical dots and/or globules

Streaks

Crystalline structures Off-center blotch Blue-white veil Peripheral tan structureless areas Regression structures (ie, scar-like areas, peppering, or granularity)

Atypical vascular structures

Increased variability in the width of the hole sizes and network lines, their color, and distribution Serpiginous interconnecting hypopigmented lines that surround irregularly shaped pigmented structures resembling elongated, curvilinear globules Multiple dots and/or globules irregularly distributed within the lesion or asymmetrically located off center or focally at the periphery; not associated with the pigment network Radial projections at the periphery of the lesion extending from the tumor toward the surrounding normal skin; may present as pseudopods or radial streaming Shiny, white linear streaks that are often oriented parallel or orthogonally to each other Off-center homogeneous areas of pigment that obscure visualization of any other structures; may be dark brown to black Hazy confluent blue white color devoid of structures Structureless areas located at the periphery of the lesion larger than 10% of a lesion area Scar-like depigmentation: well-defined areas with irregularly shaped borders and white color. It can be present in association with peppering/granularity, which consists of tiny gray dots/granules; giving the appearance of a blue-white, blue-gray, or gray-white color. The variations of the colors are determined by the density of melanin in the papillary dermis Dotted vessels: red dots of 0.01 to 0.02 mm Serpentine vessels: linear irregular or undulating short vessels Polymorphous vessels: combination of two or more vessel morphologies Corkscrew vessels: coiled, tortuous vessels Milky red globules/vascular blush: poorly defined globules with a milky red color and poorly defined areas of milky red color

Other melanoma-specific structures seen in sun-damaged skin

Periadnexal/follicular granularity Asymmetric gray perifollicular openings Angulated lines

Follicle/adnexal opening obliteration Circle within a circle or isobar pattern

Dots aggregated around hair follicles or around ostial openings Dots aggregated around hair follicles in an asymmetric fashion Multiple instances of grey or brown straight lines or pigment borders meeting at angles. Includes rhomboidal structures,20 zig zag lines,21 and large polygonal structures7 Rhomboidal structures become broader, obliterating hair follicles/ostial openings Concentric pigmented rings encircling each other